Impact of endograft design and product line on the device cost of endovascular aneurysm repair

Transcription

1 From the Society for Clinical Vascular Surgery Impact of endograft design and product line on the device cost of endovascular aneurysm repair Robert J. Feezor, MD, Thomas S. Huber, MD, PhD, Scott A. Berceli, MD, PhD, Peter R. Nelson, MD, James M. Seeger, MD, and W. Anthony Lee, MD, Gainesville, Fla Objective: Device cost is a substantial component of the overall cost of endovascular abdominal aneurysm repair (EVAR), and the four commercially available devices differ significantly in the cost of their basic configuration. This study examined the impact of three different endografts and their product lines on the overall cost of repair. Methods: Implant records of 467 EVAR procedures performed during 2000 through 2006 were reviewed. The three devices used were the AneuRx in 178 (38.1%; Medtronic, Santa Rosa, Ca), the Excluder in 123 (26.3%; W. L. Gore & Associates, Flagstaff, Ariz), and the Zenith in 166 (35.5%; the Cook Zenith (Bloomington, Ind). The Powerlink device (Endologix, Irvine, Calif) was not studied. The specific device implanted was determined by its commercial availability at the time of repair, patient anatomy, and surgeon preference. Retail list prices were used for all calculations, and only devices used during the original repair were used for analysis. Results: The device cost of the most basic configuration for repair (ie, 2 pieces for AneuRx and Excluder, 3 pieces for Zenith) differed by $3022 between the most expensive (Zenith) to the least expensive (AneuRx). However, the AneuRx system required the most number of extensions ( per case; range, 0-7), whereas the Zenith required the fewest ( per case; range, 0-3). When the costs of the extensions were added, the overall mean device costs per case were similar. Conclusion: The initial cost advantage of the AneuRx and Excluder endograft systems were offset by the more frequent need for proximal and distal extensions. The minimum device cost of a basic repair should not factor into the decision to select one specific device over another because additional devices may be required depending on the design and construction of the endograft system and the accuracy and reliability of their deployment mechanisms. ( J Vasc Surg 2008;47: ) The repair of infrarenal abdominal aortic aneurysms METHODS (AAA) has undergone a dramatic change during the past We retrospectively reviewed a prospectively maintained two decades. The traditional open surgical approach has clinical database of all endovascular repairs of AAA (EVAR) been largely replaced by endovascular techniques; however, during a 7-year period ( ). All of the procedures certain anatomic criteria must still be met to achieve awere performed at a single tertiary-care university medical successful and durable repair. The cost of these devices hascenter by one of five vascular surgeons. Preoperative planning was based on thin-cut (2- to 3-mm slice thickness) been a subject of considerable controversy regarding the overall cost-effectiveness of the therapy and remains a fiscalcomputed tomography (CT) angiograms with threedimensional (3D) reconstructions and center-path analy- obstacle to its wider adoption in certain health care systems. The current study compared the overall costs of threeses performed using one of three software platforms: Med- Metrix Systems (Hanover, NH), Vitrea (Vital Images, of the four currently commercially available devices in theical United States: the AneuRx (Medtronic, Santa Rosa, Calif), Minnetonka, Minn), and Aquarius (TeraRecon, San Mateo, Calif). the Excluder (W. L. Gore & Associates, Flagstaff, Ariz), and the Zenith (Cook, Bloomington, Ind). Specifically, we The specific device chosen was determined by its com- availability at the time of repair (AneuRx, Septem- sought to compare the costs of completing an endovascularmercial repair under actual, real-world conditions with an unselected cohort of AAA anatomies that were deemed suitber 1999; Excluder, November 2002; Zenith, May 2003), patient anatomy, and surgeon preference. The principles of endovascular repair remained relatively constant throughout the course of the study regardless of the device used and able for endovascular therapy. were similar amongst the different operating surgeons. More specifically: From the Division of Vascular Surgery and Endovascular Therapy, University of Florida College of Medicine. Competition of interest: Dr Lee is a consultant for Cook and Medtronic. Presented at the Thirty-fifth Annual Symposium of Society for Clinical Vascular Surgery, Orlando, Fla, Mar 21-24, Correspondence: W. Anthony Lee, MD, Division of Vascular Surgery and Endovascular Therapy, University of Florida, PO Box , Gainesville, FL ( leewa@surgery.ufl.edu) /$34.00 Copyright 2008 by The Society for Vascular Surgery. doi: /j.jvs The main device was routinely deployed as close to the lowermost renal artery as possible using the full length of the infrarenal neck. A proximal cuff was used if the primary device was deployed 10 mm below renal arteries or if there was 10-mm fixation on completion angiography, regardless of an endoleak. We believed that this yielded the best chances for long-term outcome. 499

2 500 Feezor et al JOURNAL OF VASCULAR SURGERY March 2008 Table I. Demographics and selected procedural measures among the three devices studied a Demographic AneuRx Excluder Zenith No. of patients Age, mean SD years Female, No. (%) 13 (7.3) 14 (11.4) 21 (12.7) General anesthesia, No. (%) 166 (93.3) 61 (49.6) 74 (44.6) Fluoroscopy, mean SD min IV contrast, mean SD ml Procedure time, mean SD min EBL, mean SD ml Urgency, No. (%) Elective 159 (89.3) 110 (89.4) 159 (95.6) Symptomatic 11 (6.2) 5 (4.1) 6 (3.6) Ruptured 8 (4.5) 8 (6.5) 1 (0.6) LOS, median (range) days 2 (1 61) 1 (1 40) 1 (0 26) EBL, Estimated blood loss; IV, intravenous; LOS, length of stay. The iliac limbs were always extended as close to theminimum cost of a repair was $12,800. Owing to the sheer hypogastric arteries as possible, regardless of common number of different diameters and lengths of the main body iliac length. In cases of ectasia (common iliac diametersand iliac limbs and extensions ($1500), the entire product 20 mm), flared limbs or bell-bottom techniques matrix consisted of 94 devices at a total cost of $367, with aortic cuffs were used. This does not reflect the recent introduction of the 36-mm main body. During the latter half of the study period, the Excluder Only the direct costs of the devices were analyzed. The flared contralateral limbs became available (October 2003), cost of ancillary devices, such as balloons, bare-metal peripheral stents, and coils, were estimated to be equivalent whereas flared limbs were available on both ipsilateral and contralateral limbs in the Zenith device since its initial among the three devices. Data are presented as mean commercial introduction. More recently, Medtronic had standard deviation where appropriate. The Student t test introduced the AneuRx AAAdvantage line extension that for continuous and the Fisher exact test for categoric variables were used for analysis. Statistical significance was included flared iliac limbs, but the AneuRx implants included in this study all occurred before this. In cases where achieved at P.05. the iliac artery was 20 mm in diameter without a suitable landing zone proximal or distal to the aneurysmal segment, RESULTS the iliac limb was extended to the external iliac artery with either surgical revascularization or coil embolization of the hypogastric artery. During the 7-year period, 467 patients underwent EVAR using one of the three devices. These included 178 AneuRx (38.1%), 123 Excluder (26.3%), and 166 Zenith (35.5%) cases. The mean age of patients undergoing EVAR Device configurations and list prices (US $) Medtronic AneuRx. The minimum construction for was years, and 48 (10.3%) were women. repair using this system is two pieces consisting of a mainpatient age among the three groups was similar, and each body-ipsilateral limb with a covered length of 13.5 andgroup had approximately the same proportion of women 16.5 cm, and a contralateral limb with a length of 8.5 and (Table I) cm. Straight limb extensions and aortic cuffs ($2025) Significantly more AneuRx patients underwent general were available. At the time of this study, the minimum list anesthesia vs local or regional, which was simply reflective price of one main body ($7250-$7550) and one contralat-oeral limb ($2225-$2425) was $9475. The total cost of thestudy period. The mean estimated blood loss, fluoroscopy the anesthetic technique used during that segment of the entire product matrix (30 pieces comprising one of eachtime, and amount of contrast used were similar among the diameter and length) was $117,500. three groups. The total procedure time was significantly W. L. Gore Excluder. Similar to the AneuRx, this is a longer for the AneuRx cohort ( minutes) than for two-piece construction consisting of a main body-ipsilateral patients who received the Excluder ( minutes) limb ($6980) and a contralateral limb ($3223). The limband Zenith ( minutes P.0001). The median and proximal extenders cost $2259. A simple two-piece length of stay was 2 days for the AneuRx group, and 1 day repair would cost $10,203. The cost of the entire inventoryfor the Excluder and Zenith groups (Table I). of 39 pieces would be $187,539. This does not reflect the The proportion of patients requiring elective EVAR flared iliac limbs introduced after the original commercial was similar among the three devices. In the entire cohort, release of the system. 450 patients (96.4%) had an intact aneurysm at the time of Cook Zenith. This is a 3-piece system consisting of a repair. The mean maximal aortic diameters ( main body ($7400) and two iliac limbs ($2700 each). Themm for AneuRx, mm for Excluder, and 59.8

3 JOURNAL OF VASCULAR SURGERY Volume 47, Number 3 Feezor et al 501 Table II. Abdominal aortic aneurysm anatomy among the three devices Anatomy AneuRx Excluder Zenith Intact, No (%) 170 (95.5) 115 (93.5) 165 (99.4) AAA size, mean SD mm Neck diameter, mean SD mm Neck length, mean SD mm Iliac aneurysm, No (%) 63 (35.6) 29 (23.6) a 53 (31.9) a Patients who had an Excluder repair had fewer concomitant iliac aneurysms (P.05). Table III. Extension usage among the three devices Extensions AneuRx Excluder Zenith Total extensions, No Proximal Distal Extensions per case Cost of extension $2025 $2259 $1500 Cost of extensions per case, mean $3857 $2681 $ or 3-piece repair, minimum cost $9,475 $10,203 $12,800 Total mean cost a $13,332 $12,884 $13,116 Cost relative to AneuRx $0 $448 $216 a 2-or 3-piece construction cost of extension. Fig 1. Histogram of extension usage for each device. Most of the AneuRx cases required two or more extensions to complete the repair, and 68.3% of Excluder cases required at least one extension to complete the repair. For the Zenith implants, 83.1% of patients required only the basic three-piece configuration, without any additional pieces mm for Zenith) and the aortic neck diameters and endograft attachment sites and junctions, whereas the lengths were comparable in all groups (Table II). SlightlyAneuRx did not. Factoring the approximate $350 cost of fewer patients in the Excluder group had a concomitant this balloon, however, did not materially alter the overall iliac aneurysm compared with the rest of the cohort (23.5% mean device cost of the repairs. for Excluder vs 33.7% for AneuRx and Zenith, P.05). To determine if the number of extensions used was Each AneuRx repair required an average of 1.90 extensions (range, 0-7). By comparison, the Excluder device the mean number of extensions used for each of the three reflective of a device-specific learning curve, we compared required 1.19 extensions per case (range, 0-4), and the devices in the first 30 patients and compared that with Zenith required 0.21 extensions per case (range, 0-3; the last 30 patients with the same device. Surgeon experience appeared to show an inconsistent effect on Table III, Fig 1). When the cost of the extensions was extension Fig 2. Mean overall cost of the endograft (cost of the basic 2- or 3-piece construction plus the mean cost of extensions) per case. included, the differences in the overall costs of the endografts for the three types of repairs were small and $500 of each other. Of interest was that although the basic cost of the AneuRx two-piece repair was the least expensive, in actual practice it was the most costly, and the Zenith system, whose basic three-piece construction was the most expensive, actually cost less than the AneuRx (Fig 2). The Excluder and Zenith endografts required obligatory ballooning with a compliant aortic occlusion balloon to mold the usage (Table IV). The number of extensions used with the AneuRx device increased significantly, whereas the number of extensions required among patients repaired with a Zenith device decreased with experience. Some of this may represent the paradox of learning curve, where more complex cases are treated as experience is gained, which in turn may require a higher number of devices.

4 502 Feezor et al JOURNAL OF VASCULAR SURGERY March 2008 Table IV. Rate of extension usage (mean SD) between the first and last 30 consecutive series of patients for each device Group AneuRx Excluder Zenith First 30 cases Last 30 cases P not hold. However, when such restrictions are lifted by operator discretion, off-label use of additional devices, such as aortic cuffs in iliac limbs, occurs frequently. It is in this context that the results of this study should be viewed. In this study, 3D reconstructions of CT angiograms were used routinely for preoperative planning and sizing of the cases. Conventional preoperative angiography with marker catheters was rarely used. Although diameter measurements have become fairly reliable, path-length measurements remain the least reliable indicator in endograft sizing and planning, regardless of which method is used. At follow-up, reinterventions were required in 29 Axial CT tends to underestimate the path length, 3D AneuRx (16.3%), 13 Excluder (10.6%), and six Zenith reconstructions tend to overestimate, and angiography falls (3.6%) patients. For those late secondary procedures that somewhere in between, depending on a host of anatomic required placement of extensions, more proximal extensions were used for the AneuRx device (n 13) than used to compensate for these variabilities in the different factors and techniques. In practice, fudge-factors are Excluder (n 0) and Zenith (n 1) devices. Similarly, imaging modalities. One of the main indications for additional extensions in the AneuRx and Excluder cases in this distal extensions were required more frequently with AneuRx (n 6) than Excluder (n 3) or Zenith (n 0). series was to achieve limb extensions to the hypogastric It should be noted, however, that reintervention is a time-arteriedependent event, and the AneuRx cohort had the longestwas purposely selected in borderline cases to avoid inadver- when a shorter main body and contralateral limb mean follow-up among the three groups. tent coverage of the hypogastric artery. Arguably, had an accurate and reliable method of DISCUSSION length measurement been available, iliac extension usage This study examined the real-life costs of different could have been decreased. Alternatively, a method of endograft systems used to treat an unselected, consecutive trombone-ing the iliac limb within the docking gate of series of patients with AAA suitable for EVAR. Our results the main body would allow sufficient intraoperative flexibility to overcome this preplanning deficiency. This con- showed that the basic cost of repair meaning the minimum number of devices required to achieve a bifurcated cept of intraoperative adjustability of the iliac limbs can be construction does not tell the whole story. Indeed, despite an initial cost differential of $3000 from the least both limbs to provide the maximum degree of deployment extended not only to just the contralateral limb but also to expensive device to the most expensive device, in the final flexibility. analysis, the mean cost of actual repair differed by $500 Another method of potentially reducing extension usage would be, theoretically, to stock an unlimited inventory among the three devices. The reasons for this include what is considered optimal endovascular repair, endograft-dependent factors, on-the-fly. Even with wider availability of consignment where intraoperative selection of any device can be made and the current limitations of sizing and planning. In our stock by the different manufacturers, it is generally impractical from the standpoint of space and inventory manage- practice, we strongly believed that optimal repair is achieved with maximal coverage of the proximal and distal ment to stock the entire product matrix. A more realistic landing zones. Although there is little controversy about alternative is if multiple lengths of the same diameter devices could be ordered without additional financial burden deployment of the main body as close to the renal arteries as possible, 2 there is some controversy about the routine (per-case consignment) so that the operator has the option 3,4 extension of the iliac limbs to the hypogastric arteries, of choosing the optimal lengths of the devices at the time of especially when there is 2 cm or more of relatively undiseased or nonaneurysmal artery. As reported by Arko et al, repair. During the period of this study, such purchase the procedure and pay for only those devices used for 5 for the AneuRx system, which relies on columnar support agreements were not available, and ordering two main as an important mechanism of fixation, it has been shown bodies of differing lengths was not a fiscally viable practice. that full coverage of the common iliac arteries resulted in On the other hand, the capacity to have some method of better late outcomes in terms of migration and limb retraction. Furthermore, there is a small but definite risk of late amount of inventory one would have to stock. Along these adjusting length built into the device would help in the aneurysmal degeneration of the uncovered iliac segment. lines, a larger variety of device configurations would further Once commercialization of a medical device occurs, obviate the need for aortic cuff usage for bell-bottoming application of that device in anatomic situations outside of purposes. Although not applicable for this study, flared iliac the manufacturer s instructions for use (IFU) is common. limbs are now available for all the devices examined in this This is obviously in contrast to the strict constraints of article. clinical trials that comprise the initial experience of any new Deployment accuracy also affects cuff and extension medical technology. In the idealized conditions of a clinical usage. Especially proximally, in cases of neck angulation, trial and strict guidelines of the device IFU, our results may different mechanisms of deployment can significantly influ-

5 JOURNAL OF VASCULAR SURGERY Volume 47, Number 3 Feezor et al 503 ence the need for a proximal cuff to remediate a low, among many other equivalent or potentially even more suboptimal deployment. Indeed for AneuRx and Excluder valid surrogate measures. Arguably, in endovascular repair, cases, proximal cuffs accounted for 31% and 32%, res-thpectively, of the entire extension usage (vs 0% for Zenith; of the immediate cost of the therapy because other factors cost of the endograft comprises a dominant component Table III). such as use ancillary devices, routine postoperative care, and The study has a number of limitations. First, as men-lengthtioned, the more recent expansions to the product lines of of EVAR is concerned, clearly, many other factors must be of stay are all fairly similar. As far as the overall cost the respective devices were not considered. Most of theconsidered such as follow-up imaging and the cost of experience reported here predated the additional sizes and reinterventions, which speaks more to the long-term efficacy of a particular device. lengths that could have materially impacted the overall results and their applicability. Second, although only the endograft costs were considered, there may be other ancillary disposable equipment The basic cost of repair using a particular device does CONCLUSION unique to a particular endograft system that could impactnot always equal the actual cost. In fact, the overall costs of the overall device-related cost of the procedures. However, the devices were fairly equivalent among the three devices any variability would likely have been a small fraction of the considered. Therefore, the minimum cost of the two- or overall cost. three-piece construction should not influence the decision In a related manner, this analysis was based on a static to use a particular endograft system from a fiscal standpoint. cost of the devices. Were those prices to change signifi-cantly, our results might be significantly altered. To ourment of lengths may reduce need for additional devices and large product matrix and ability for intraoperative adjust- knowledge, although the AneuRx device has been discounted slightly from its original list price, no substantial the final endograft-related cost of repair. price changes have occurred in any of the other endograftauthor CONTRIBUTIONS systems. Given the current competitive marketplace, however, creative consignment and purchasing programs indi- Conception and design: RF, WL Analysis and interpretation: RF, WL vidually negotiated between the hospital and the vendor are Data collection: RF, WL more widely available today than in the early years of this Writing the article: RF, WL study. This could affect the ability to intraoperatively Critical revision of the article: RF, TH, SB, PN, JS, WL change device selections and lower the need for extensions. Final approval of the article: RF, TH, SB, PN, JS, WL Third, our bias of routine iliac extension to the hypogastric arteries is not supported by any clear scientific Statistical analysis: RF, WL Obtained funding: Not applicable evidence of improved patient outcome. If such a policy had Overall responsibility: WL not been practiced, the results could have been significantly different than reported. Fourth, complexity of an endograft system (2-piece vs REFERENCES 3-piece construction, single-stage deployment vs multistage deployment) may lead to prolongation of procedure et al. Bell-bottom aortoiliac endografts: an alternative that preserves 1. Kritpracha B, Pigott JP, Russell TE, Corbey MJ, Whalen RC, DiSalle RS, time and increased operating room costs. Of interest, however, was that the mean procedure time was almost 602. Lee WA, Zairns CK, editors: Infrarenal abdominal aortic aneurysm: pelvic blood flow. J Vasc Surg 2002;35: endovascular repair. St. Louis, MO: Mosby Inc; p minutes longer for the AneuRx procedures than for the 3. Hobo R, Laheij RJ, Buth J. The influence of aortic cuffs and iliac limb Excluder and Zenith procedures. This may be a reflection extensions on the outcome of endovascular abdominal aortic aneurysm of a learning curve effect and the additional time expended repair. J Vasc Surg 2007;45: for implantation of the extensions. Finally, we did not consider the Powerlink (Endologix, Irvine, Calif) device, a unibody device that may offer a significant competitive fiscal advantage compared with the other three devices. In this study, we arbitrarily chose a rather narrow measure of device cost as a measure of the cost of repair 4. Karch LA, Hodgson KJ, Mattos MA, Bohannon WT, Ramsey DE, McLafferty RB. Management of ectatic, nonaneurysmal iliac arteries during endoluminal aortic aneurysm repair. J Vasc Surg 2001;33:S Arko FR, Heikkinen M, Lee ES, Bass A, Alsac JM, Zarins CK. Iliac fixation length and resistance to in-vivo stent-graft displacement. J Vasc Surg 2005;41: Submitted Jul 17, 2007; accepted Oct 27, 2007.